TACTICAL ISSUES

We have already alluded to Kantor's conception of the psychological event as a requisite foundation for behavioral research. We also noted the fact that Kantor included the investigating scientist as a field factor. Figure 14 illustrates all attending field factors from this point of view. These include: the observing scientist, the active organism, the environment, and the temporal setting conditions of past, present, and projected future conditions. This model, in fact, establishes our position on two important issues: (a) how behavioral systems shall be defined, and (b) the role of the investigator in the system. We briefly turn to the implications of our stand on these issues.

Role of the investigator. The primacy of observational events as phenomenologically experienced by the observer (i.e., as empirically established via the observer's sensory capacities and/or instrumented extensions of same) is always represented to the scientific community only secondarily via the observer's descriptions of the events. Although replication may allow us individually to experience similar primary events ourselves, the real stuff of science is in the communicated measurement (i.e., description) of events observed, not in the events themselves. All measurement essentially begins with verbal descriptions, even if descriptions are eventually refined into more precise terms, such as quantitative measures. As such, systems methodology, like previous methods, requires close scrutiny and tight, if sometimes different, rule governance regarding how observations are to be selected, recorded, and communicated. In fact, describing the observer's role in the system is as crucial a specification in behavioral systems analysis as in the description of any other component of the system (Ray, 1977). This, of course, suggests the need for a rather elaborate descriptive system for organizing the varieties of observer behaviors, as well as those of the observed system (cf., Boice, 1983). That is yet another purpose of the present survey of systems strategies and tactics: to systematize heretofore poorly organized methodological nosology (i.e., descriptive theory) regarding the analysis of behavioral systems.

It should be clear by now that all forms of description are actually symbolic representational systems which replace the primary events being described. Most of the time these symbolics emanate exclusively from the observer and do not include descriptions by the primary subject in the behavioral system. Classical phenomenology stresses the importance of comparing observer descriptions of events with those of the participants as well. Thus it may be argued that a subject's descriptions reveal, at one level, some rules of the primary system perhaps even better than descriptions by the observer, because the observer's descriptions are governed more by externally prescriptive (i.e., methodological) rules, rather than organizational rules governing the internal systemics. Eventually, the strength of any methodology lies in its ability to assure a maximal coupling between primary events being described and the events of actually describing. Such accuracy of description, of course, is a function of discrimination training in the use of appropriate language (a historical setting factor) and may apply to scientist and subject alike.

Informal linguistic descriptions. When human observers experience a unique phenomenon, it seems almost inevitable that they translate various aspects of their experience into a language-based description. Whether they describe it subvocally to themselves, talk to others about it, or write about it, if the impression is lasting, they most often convert it to words. In circumstances where several others experience the same phenomenon, a "localized cultural" (i.e., socially shared) language is often formed for communicating aspects of the event. By this we mean that a standardization of terms often occurs whereby socially shared experiences are given singular labels. These labels reduce the complexity of original, almost groping, descriptions which are more lengthy and convergently redundant, and they aid communication where uncertainty of reference is a possibility. Thus, as anyone knows who has ever developed a more formal descriptive research category system, formal categories typically are not where descriptive research starts, but rather represent a stage of eventual refinement. What gets refined are the virtually naive descriptions which often mark the pilot and/or qualitative stage of descriptive research.

Cultural anthropologists and field primatologists are perhaps best known for applying informal linguistic procedures for description. They collect reams of field notes for later systematization and/or formal interpretative analysis (cf., Trend, in press). As such, these scientists share with all other scientists the establishment of their own technical, descriptive, and closely shared cultures via their generation of jargonistic terms to represent and compare the various events under study. Whether they recognize it or not, all researchers, not just field scientists, begin research at this informal linguistics level-if their investigational system is relatively new to scientific investigation. Therefore, it is relevant to consider the character of informal linguistic descriptions on a broader scale.

For example, before Ray et al. (1977) began to generate the formal categorical descriptions of killer whale behavior cited earlier (see Figure 3), they had many informal conversations with animal trainers who shared similar histories of experience with this species. These trainers are examples of the localized culture of descriptive language that emerges around closely shared and unique phenomenal experiences. For example, trainers will often, with apparently precise communication among themselves, speak of an individual animal as being "squirrelly" during show performances. Because the scientific literature on learning/performance maintenance offers little to enlighten such references, Ray et al. assumed a new phenomenon might well be implied. As with these researchers, without lengthy discussions and convergently defining descriptions as guides, readers may assert their own naiveté regarding animal training culture and may acknowledge sharing little direct experience in identifying what "squirrelly" means either. Most readers would hardly be expected to know a "squirrelly" animal from a "normal" one. Nevertheless, they might well jump to analogy building and assume that the behavior of killer whales might occasionally mimic those of squirrels in form, speed, predictability, etc., and thus could easily justify the translation from the noun "squirrel" to an adverb describing "squirrelly actions." Only further, and highly analytic, participation with the describing culture really allows such hypotheses to be confirmed and/or rejected. As it turns out, squirrelly appears to be a term relating to an animal's likelihood of responding to discriminative cues with inappropriate, unpredictable, and often fleeting behavior just prior to carrying through with what is eventually appropriate behavior to the command signal.

Informal linguistic descriptions rely heavily upon a certain naive and/or unstudied substitution of language for primary events. But closer inspection of this process reveals a variety of implicit problems which require further assessment. Because informal descriptions, whether by subject or observer, typically start with this localized cultural language and usually take the form of multiple complete sentences, a formal analysis of informal descriptions relies upon relatively standard procedures typical of formal linguistics and/or textual analysis (Hockey, 1980; Oakman, 1980). By analyzing observers' and subjects' sentence structures and domains of reference, including their inclinations to use personal pronouns versus nouns, their use of adjectives and adverbs as qualifiers, their usage of verbs and verb tenses, and their incorporation of spatial and object references, as well as relational terms as reflected in prepositions and other autoclitics, we can begin to discern precisely what about the system under investigation is prompting the observers' descriptive attention.

For example, a study by Ray, Griffiths, and Wruble (1987) explored the hierarchy of naive observers' inclinations to use different domains of descriptions when observing a typical behavioral data source (baboons on videotape) for the first time. They further assessed the implications of multiple viewings and the impact of still versus moving pictures as sources of events to be described. They found that the majority of naive subjects' descriptions included references to both the action of the baboon subjects and the objects upon which that action was directed. Yet a sample of 37 categorically based professional studies on baboons (Griffiths, 1987) demonstrated almost no formal categories for including specific objects of behavioral interaction in the description, other than social objects (i.e., other baboons). Interestingly, this inclination for naive observers to focus upon action-object descriptions is more consistent with the behavioral systems approach as herein defined than is the majority of the professional primatology studies reviewed by Griffiths. But it is also clear from the Ray et al. (1987) study that the temporal demands imposed by events on observers are most often ill suited to linguistic reactions. Complete sentences, which typify informal linguistic descriptions, tend to be slowly constructed relative to the typical speed of ongoing behavioral interactions. Like a football game being followed via descriptions over the radio, most of the action details are lost and only highly selected aspects are described. Many elements may be skipped over, more as a matter of keeping pace than as one of reflecting the primacy of those events that are selected for description. Such temporal demands often lead researchers to utilize supportive technologies such as video recordings and to slow, freeze, or even replay event sequences for the advantage of more encompassing descriptions. Nevertheless, multilayered events frequently remain problematic in informal linguistic descriptive systems.

Unfortunately, informal-to-formal linguistic translations are rarely discussed or illustrated in the scientific literature. Systematic methodological research at this level is hardly deemed sufficiently important by most to warrant publication, and few journal policies support manuscript acceptance reporting such research. However, the systems perspective suggests that it is important to better understand this process of refinement (i.e., the selectivity in retention/rejection of descriptive domains and categories) which marks transitions from informal description to formalized category systems. Ray and Wruble (1986) conducted empirical research on this neglected question. (We gratefully acknowledge The MacArthur Foundation for financial support and Dr. Orville Smith and his colleagues at the Regional Primate Research Center, University of Washington, Seattle, for making available the physical and collegial resources which made this study possible.)

The videotape was subsequently replayed both in real time and in slow motion while Ray and Wruble discussed each action sequence and its attendant properties. This discussion culminated in a descriptive narrative account that was subsequently used to determine key domain references and probable categories detailing these domains. This involves a form of phenomenological -linguistics analysis that probes for generically related domains of descriptive reference (including domains of behavior, location, object involvement, directional orientation, social individuals supporting interaction, etc.). This step was used to identify all potentially relevant domains for which exhaustive categories needed to be developed and operationally defined. This also supplied a relatively detailed, albeit preliminary, list of such categories as well. This preliminary categorization system was then used to transduce the entire three hours of videotaped data. This process identified any categories and/ or domains potentially missed in the first stage, but still considered relevant to the study's primary purpose.

The next step toward formalization involved application of the preliminary category system to the analysis of approximately three hours of videotaped data focusing on another subject in this same group-the subordinate (beta) male. A coded transcription of this tape generated still further confirmation and refinement of the domain/category system. This eventual product seemed, at least to other trained researchers familiar with the species and intention of description, highly adequate for the task of reconstructing all vital elements of any given interaction by any subject, whether under isolate or social circumstances. Table 2 gives a summary of this category system and reflects the retained domains as well as the inclusive descriptive categories necessary for describing the elements of that domain. Thus actual scoring, following the domains listed in Table 2, began with a notation of the exact time, in terms of hour, minute, and second (HMS), on the videotape that a change in any domain actually occurred (column 1). Second, any relevant spatial objects and spatial zone were noted (column 2 and 4 respectively). The subject's direction of orientation, postural activities, head activities, etc. were noted (columns 3-5-9), in addition to a subject's relationships (column 10) with physical/social objects (columns 12-13). Column 11 notes any physical versus zone types of contact with social objects (i.e., other baboons).

TABLE 2. First-stage Category system Developed for Describing Baboon-environment Interactions by Ray and Wruble (1986).

The coded narrative from this analysis was then subjected to empirical "linguistic reconstruction," or transliteration analysis, which represents the more unique and empirically oriented efforts to understand the process of linguistic formalization. A computer program, following standard grammatical and syntactic rules of sentence construction, reconstructed a narrative-based verbal transliteration of the coded data file, including all elements involved in each consecutive event recorded. Table 3 gives an illustration of only a brief, but relevant, segment of the computer file format of these coded data. In this table, the domains of Table 2 have also been coded into a two-letter heading, and the categories are coded according to the capitalized letters in each as depicted in Table 2. Only those domains that changed are entered, and the stable categories are merely represented by two periods (thus noting a continuation of that state). Table 4 offers the computer-generated transliteration of the data in Table 3 and illustrates how the coded data are formed back into narrative English sentences. The intent here is to translate a simple two- letter coded sequence back into a clear and understandable sentence which reflects the code's full meaning and relationship with other domains. Thus the first coded line of Table 3 begins with BM as the subject's identification code (ID) and the time (00 39 16 86) to the nearest thousandths of a second on the digital time code generator. Next, the subject's zone (A5) and direction (S) are given, etc. The transliteration of this information replaces BM with "The focal subject for this description is the beta male, Alfred," and the A5 zone with "is located along the north fence/wall near the middle of the shelves" (Figure 4, lines 1-4). The "s" code translates into "and is facing south" (line 4). The remainder of the file is similarly decoded.

This analysis subjects the coding system to a complete assessment regarding precision of linguistic syntax and the existence of potential reference ambiguities. Ambiguous or missing references become quite obvious from such an analysis, as evidenced in lines 23 through 26 of Table 4. The text describes a subject who is shifting from being located near another subject to the position of being above himself! In addition, he is identified as receiving social contact with himself. Both of these are totally in violation of reality, but they are logical extensions of the way categories within certain domains (i.e., relational, orientational, and physical contact) normally relate to categories of other domains (i.e., objects and animals as social individuals). It is grammatically correct to have some nouns as objects of prepositions and as predicates, but others function somewhat more exclusively. For example, self and other animals represent such an exclusionary distinction. This is not always obvious until a relatively formal analysis of the implicit linguistic relations among categories and/or domains are actually studied empirically and their reflection of reality confirmed. Formal linguistic/stylistic analysis thus serves as one for of "quality control" in the formalization of categories which derive from informal linguistic origins.

Other problems arise with categorization of informal linguistic descriptions as well. For example, researchers highly schooled and/or experienced in category-building technologies often identify with problems of appropriate (and consistent) "levels of resolution" between categories, consistency of domains of reference (cf., Purton's 1978 discussion of structural/functional conflation and Schoenfeld's 1976 discussion of the "response" in behavior theory), and even problems of boundary clarity. Such problems, while being tactical in nature, unfortunately are too complicated to address here. Nevertheless, these problems clearly are related to issues of linguistic formalization and referential consistency.

Formal linguistic systems enable researchers to observe and record several aspects of a phenomenon nearly simultaneously and hence overcome the linearity of informal language. This is the approach embodied in musical notation systems, where specific aggregates of notes are both nominally ("A" versus "C") and ordinally represented (i.e., musical notations indicate not just high/low notes, but also tempo, intensity, etc., all of which are identifications of a variety of ordinal domains of parametric note variations). Nevertheless, even formal linguistics are often found lacking in power and application. Some observers recognize the limitations of linguistic descriptions and record events by other representational strategies, many of which may be identified as demonstrative description.

Demonstrative description. Although the simplest form of demonstrative description is the direct pointing to events as they occur in socially shared situations (or the restaging/replication of those events), perhaps the more familiar demonstrative description in science is the use of drawings of the original phenomenon and/or summary graphics (Tufte, 1983). Musical scores, as noted, symbolize both instrument layout and intended playing sequences (including the temporal characteristics required) by the use of a special kind of graphic plotting system to guide musicians' behaviors. More detailed, and representational, graphics were used as description as far back as the invention of hieroglyphics. Leonardo da Vinci perhaps saw the greatest utility in wedding the graphic arts to science, as his copious notebooks will attest. And the Dutch Renaissance painters of the 16th century brought pictorial representation as description to its near epitome (cf., Alpers, 1983) well before photography was invented to serve the same purpose. Later, impressionistic painters preferred to abstract the events they witnessed, reducing them only to essential and/or emphasized form and color relationships. Movement was finally added to the pictorial descriptive arts via cinematography which had a very immediate and dramatic scientific impact beginning with Muybridge's (1887) recordings of human and animal locomotion.

Perhaps the more frequent utilization of emotive description in scientific communication appears at professional presentations, rather than in the printed media, where external editors can caution, cajole, or otherwise diminish the frequency of such descriptions. In oral deliveries, the words can be given a spontaneous (or not so spontaneous, in a great many cases) intonation, inflexion, and amplitude which characterizes the truly emotive domain. The teaching of science is far more encouraging of "enthusiastic" than "dry" descriptions of events. But that is because listener/consumer interest in the description is more clearly recognized and valued in the classroom ( i .e., with newly aspiring scientists) than in the scientific meeting rooms (i.e., with established scientists), and even less so in the scientific journals.

Emotive description eventually defines the Zeitgeist factors of the scientific enterprise and is indirectly indicated in many instances as much by what is done to avoid dissemination as what is done to champion a cause. Thus, Darwin's case of shelving treatise on natural selection until prompted by Wallace's manuscript tells us much about Darwin's descriptions of his findings and interpretations vis-à-vis social pressures of his time. Clearly, as any clinical psychologist will quickly attest, there is a logic (i.e., rule structure) suggested in the dynamics of emotive description. That logic reaches to the very foundation of what is, in the lay language, described by the terms "attitudes," "connotations," "feelings," and "values." Pavlovian conditioning is a primary process in establishing and generalizing such descriptive responses, but the mixture of emotive (connotative) and denotative meaning in language vastly complicates the tracking of the full generalization phenomenon. Thus we wish to suggest that emotive-descriptive behaviors are always with us in one form or another and that we should strive better to understand their rule structures, not only in scientific behaviors but in all behavioral systems. We will best do so by recognizing such behaviors and describing them within their systemic context. We also strongly suspect that the demonstrative-emotive organizational rules are somewhat different than those which govern demonstrative-verbal and demonstrative-artistic descriptions, as any drama director will quickly attest. Each has its own place, in that each is always evident. But each is not as equally recognized, described, and utilized by design and intent.